Extracting Keywords of UAVs Wireless Communication Protocols Based on Association Rules Learning

摘要

Unmanned Aerial Vehicles (UAVs) are increasingly popular. As a result of the tremendous number of UAVs, especially recreational UAVs, regulation becomes a challenge we are confronted with. Protocol reverse engineering offers a way to understand and regulate the drones effectively. Extracting keywords is an indispensable fundamental work for protocol reverse engineering. However, to the best of our knowledge, there are few works in the literature that openly address the issues of extracting keywords of UAVs communication protocols. In this paper, we propose a framework to extract keywords from the binary data collected from wireless communication channels of UAVs. Firstly, we capture the binary streams and segment into frames. And then we find frequent sequences utilizing multi-pattern matching algorithms. After that, we extract keywords based on frequency and analyze the association rules among these key words. Our experimental results show that 8-bit sequences such as 0x20, 0x60, 0x01, 0x09, 0xFF, 0x0A, 0x0D, 0x7D, 0x1F, 0x51, 0x49, etc. are keywords. And our framework identify frames of UAVs with a precision rate about 88% and a recall rate of around 87%.